Multiple sclerosis (MS) is a common and severe disorder of the central nervous system characterized by chronic inflammation, myelin loss, gliosis, varying degrees of axonal and oligodendrocyte pathology, and progressive neurological dysfunction. MS pathogenesis includes a complex genetic component. In spite of intensive long-standing efforts, the knowledge of MS genetics remains incomplete and significant genetic questions still remain unanswered. In particular, we plan to address how susceptibility variants influence disease susceptibility and whether variants also influence disease trajectory. This application is primarily hypothesis and data-driven, and will employ a combination of advanced bioinformatics and experimental approaches thus bridging efficiently the dry and wet laboratories to advance our knowledge on the pathways at play in MS. The unifying rationale of this proposal is to functionally link DNA polymorphisms (i.e. at the bottom of the hierarchical organization of biological complexity) with higher strata (e.g. protein networks, imaging endophenotypes) through the integration of information from intermediate levels. We propose these goals: Specific Aim 1. Cell specific eQTLs in multiple sclerosis. Building on the hypothesis that allele-specific RNA expression levels can act as surrogate of the functionality of the risk variants, we will integrate the most updated SNP-based MS dataset with information available on the encyclopedia of DNA (regulatory) elements (ENCODE) and additional databases in an effort to identify robust eQTLs and narrow the roster candidate causative variants. We will study their functional relevance for MS by RNAseq in 4 relevant cell populations (CD4, CD8, B cells and monocytes) isolated from MS patients with extreme disease trajectories. Specific Aim 2. In-silico functional analysis of gene x gene interactions. Here we propose to extend our earlier work and develop limited stratification schemes to identify gene networks enriched in well-defined extreme phenotypes (e.g. aggressive vs. benign disease). We will follow a computational strategy that takes into account whether risk alleles fall within functional elements and integrates them in a combinatorial fashion with available cell-specific gene expression datasets. The ultimate goal of this aim is to refine the MS genetics map and identify the source of genetic heterogeneity in MS. Specific Aim 3. Genotype-MRI phenotype associations in multiple sclerosis. We will examine an ongoing prospectively ascertained longitudinal cohort (n > 500) of deeply phenotyped patients for genotype-phenotype associations. This unique cohort will be studied to determine whether genetic variants genome-wide correlate with direct and indirect metrics of disease activity and progression as determined by brain MRI and optical coherence tomography (OCT). Specifically, relapse-related activity and brain neuronal loss will be estimated by global and regional measures of grey matter volume changes. We will also use OCT prospectively to measure thickness of the retinal nerve fiber layer and macular volume longitudinally as surrogates of disease activity.